Source: docs/index.md

Docs Home

Start here when the repo feels like a pile of good pages but you need a path.

This curriculum has two jobs:

Choose a tutorial

Tutorial	Use when	Shared materials
Builder Loop Tutorial	You want one real project slice from intent through execution, review, and learning.	intent, model fit, context, framing, evidence, brief, review, extraction
Context to Agent Tutorial	A useful prompt or checklist might become a reusable interface.	context construction, prompt assembly, skills, subagents
Eval Tutorial	You need a quality bar, fixture set, harness check, grader, threshold, or production signal.	evidence, eval checklist, SLO-style signals, review

The full module-by-module map lives in Curriculum. Use this page to choose a path, then open shared materials when the tutorial asks for them.

Area	Pages
Grounding	Intent Engineering, Model-Fit Framing, Context Construction, Prompt and Context Assembly
Framing	Open Horizons Phase Skills, Problem Space, Problem Statement, Beyond the Nearest Peak
Evidence and delegation	Evidence and Evals, Agent Briefs, Strategy Clarity
Reusable interfaces	Authoring Skills, Authoring Subagents, Context to Agent Tutorial
Execution and learning	Artifact Contracts, Execution, Review, Dissent, and Salvage, Knowledge Extraction
Sources	Further Reading

Need	Template
Know what each artifact must preserve	Artifact contracts
Choose a project slice	Builder playground
Supply selected context	Context pack
Assemble a prompt	Prompt assembly
Define checks	Eval checklist
Hand off execution	Agent brief
Preserve a procedure or role	Project skill, Subagent
Preserve learning	Knowledge artifact

Symptom	Go here
The model output is fluent but generic.	Model-Fit Framing, Context Construction, and Prompt and Context Assembly
You keep copy-pasting the same prompt or checklist.	Context to Agent Tutorial and Authoring Skills
Nobody knows what counts as correct.	Eval Tutorial and Evidence and Evals
The repo has too many possible fixes.	Problem Space, then Problem Statement
The first patch looks plausible but shallow.	Beyond the Nearest Peak
The agent keeps wandering.	Agent Briefs and Execution, Review, Dissent, and Salvage
The patch works, but the next session would forget the lesson.	Knowledge Extraction
The work keeps expanding and the finish line keeps moving.	Execution, Review, Dissent, and Salvage